1) The study analyzed how age, gender, personality traits, and language use on Twitter can predict and provide insights into mental illnesses like depression and PTSD. 2) The results showed that mentally ill users tended to be higher in neuroticism, more introverted, and less agreeable, and that controlling for age and gender was important. 3) Language models that incorporated a large number of features including personality, affect, topics, and n-grams were most accurate at detecting mental illnesses, demonstrating the importance of personality in understanding conditions like depression.